DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
UNIVERSITY OF NEVADA, RENO - MS 171
RENO, NV 89557-0171
Artificial intelligence with focus on computer vision is my primary research interest. Various scientific fields including robotics, medical imaging, security systems, and so many others employ the results of computer vision processes. There are many challenges faced by researchers in this field which emphasize the need for elaborate collaborations with other disciplines to develop new and to enhance the existing theoretical and algorithmic frameworks.
As an electronic engineer and a computer science graduate, I am interested in investigating various theories and their applications to real world problems. My deep interest in computer vision and its numerous subdivisions stem from the fact that it is a crossroad of science and technology. Within the area of computer vision, my specific interests include human activity/intent recognition, video processing, active vision, robotic vision, and biometrics. I believe that the versatility of computer vision and its closeness to other disciplines, through sharing of theories and applications, also helps me build strong collaborations with scientists in various division and subdivisions of the college of engineering including artificial intelligence, computer graphics, computer games, robotics, electronics, etc.
Background and Current Research
My fascination for mathematics and physics led me to the Electrical Engineering Department of Sharif
University of Technology. As a graduate student in the image processing laboratory, I developed a framework to extract different video object planes from videos for a distant learning application. I proposed to compress the extracted video object planes with different resolutions and to transmit them with different frequencies over a low-band network to improve the quality of service.
I am currently completing my doctoral degree in the computer science at the University of Nevada,
Reno. My primary research project deals with tracking of multiple objects in videos with complex backgrounds, also called quasi-stationary. I developed a scene independent, non-parametric object tracking framework that detects and tracks foreground objects while disregarding background changes. After detection of foreground regions, the system finds relevant objects and tracks them. In order to track objects in real-time, I developed a spatio-spectral connected component processing mechanism which uses object appearances to assign them unique IDs. Although my primary research has been in computer vision and object tracking in videos, I have collaborated with the robotics research group as well. My real-time object tracking algorithm has been successfully employed in a project which uses robots in an environment to recognize the intent of other agents (i.e. robots and people).
Research is an evolving and endless process. I aspire to develop a computer vision research group, to advise graduate and undergraduate students, and to collaborate with scientists and researchers within and out side the department and the school. My future plans are to improve and to expand my current research in object tracking and to explore other research areas within the computer vision field. To improve the quality of object tracking I would like to explore applicable artificial intelligence tools. Currently, the stateof-the-art object tracking systems use global motion or background models to detect the foreground. In some applications, such as traffic monitoring, certain categories of objects are of particular interest. I would like to investigate pattern recognition ideas to model classes of objects of interest for localization and tracking purposes.
Since interdisciplinary collaboration both enriches our current achievements and opens windows to new possibilities, I would like to investigate the applications of my